Are Heinrich events triggered by the ocean? From a conceptual model to a 3D ice sheet model

A common explanation for Heinrich events consists in an internal thermomechanical feedback of the Northern Hemisphere ice sheets allowing the occurrence of periodical large ice surges into the ocean. It has been proposed that these large-scale surges likely occur when basal ice reaches the melting point in regions where subglacial sediment allows very strong sliding. Attempts to simulate such mechanism (so-called bing-purge) in the framework of the shallow ice approximation (SIA) indeed succeeded to produce oscillations (HEINO experiments). However these results are still controversial because the oscillation periods were very dependent on numerical methods and, more important, because the SIA does not take into account longitudinal stresses and is thus not suitable when there are strong gradient in velocity. Using a 3D ice sheet model, GRISLI, in which ice streams velocities are obtained with the MacAyeal equation and thus longitudinal stresses are accounted for, no oscillations can be obtained with only the ice sheet-ice stream system. The recent breakup of some ice shelves in Antarctica and the associated changes on ice streams velocities illustrates the importance of the buttressing effect on ice surges. We suggest here that ocean variability has to be considered as a potential triggering factor via the ice-shelf erosion and its consequences on ice streams. With a conceptual ice sheet /ice shelf /ocean model we demonstrate the ability of this simple model to produce physically based oscillations compatible with the signature of Heinrich events that can be only generated in presence of ocean millennial variability which is suggested by all the data available.